Mystery in board assembly SOLVED
T Chap
Posts: 4,198
After 45min with AD tech, I discovered that I connected my power supply wrong for the first time in 7 years. I was feeding the ref voltage to the AD628 with 28V instead of 5V. Easy fix but 15 hours to find it the mistake.
I have made many hundreds of a BLDC motor driver board. I use the neoden TM245 PNP. The schematic shows an AD628 current sense device, which is a programmable output that reads current across two 1.2 sense resistors (R14, R10) in parallel, the inputs of 27VDC to the motor driver board. I made boards monday and tuesday no problem. Yesterday I made 6 boards and all have the same exact problem. I have only changed out 1 reel this week coincidentally, which as a 33 ohm 0603 series gate driver resistor to all 6 of the IRF540 mosfets. I verified the new reel does have 33ohm parts, but these parts are not even part of the schematic where the issue is. I need to solve this fast and often just posting here solves it mysteriously.
The output of the AD628 should be .235VDC when the motor is idle. It will hit maybe 2V under max load. The output of the AD628 feeds a LM393 comparator. The Propeller DAC( .1cap and resistor) feeds the other input on the 393 comp and sets the trip level for the current overload.
The output on all 6 boards built yesterday and today is .590V. Almost double and I am getting trips just sitting idle.
I need to get some ideas on where to start. Here is where I am:
1. Verify the parts being called on the machine did not somehow mysteriously change in the file. Normally, you cannot change a value(reel) without going into edit, go to a part, update it's stack number, save. Not sure how this can even happen accidentally at all.
2. Verify the existing reels values to confirm no manufacturing flaws on part values, ie someone the reel changed resistance values on yesterday. EDIT ALL parts on the board seem to be correct with schematic.
3. Find an old AD628 in the graveyard box and stick it in and verify there is no mfg output change on newer devices. Makes no sense there would be mfg changes in the same tube of parts though. EDIT NO OLD PARTS FOUND to test.
4. Verify the input sense R's are correct at 1.2 ohm each. EDIT Done.
In the output of the AD628 amp, there is a divider section(R18, R21). Something would have to be really off on the math to show this much difference.
Any suggestions on how to diagnose this weirdness?
Thanks
I have made many hundreds of a BLDC motor driver board. I use the neoden TM245 PNP. The schematic shows an AD628 current sense device, which is a programmable output that reads current across two 1.2 sense resistors (R14, R10) in parallel, the inputs of 27VDC to the motor driver board. I made boards monday and tuesday no problem. Yesterday I made 6 boards and all have the same exact problem. I have only changed out 1 reel this week coincidentally, which as a 33 ohm 0603 series gate driver resistor to all 6 of the IRF540 mosfets. I verified the new reel does have 33ohm parts, but these parts are not even part of the schematic where the issue is. I need to solve this fast and often just posting here solves it mysteriously.
The output of the AD628 should be .235VDC when the motor is idle. It will hit maybe 2V under max load. The output of the AD628 feeds a LM393 comparator. The Propeller DAC( .1cap and resistor) feeds the other input on the 393 comp and sets the trip level for the current overload.
The output on all 6 boards built yesterday and today is .590V. Almost double and I am getting trips just sitting idle.
I need to get some ideas on where to start. Here is where I am:
1. Verify the parts being called on the machine did not somehow mysteriously change in the file. Normally, you cannot change a value(reel) without going into edit, go to a part, update it's stack number, save. Not sure how this can even happen accidentally at all.
2. Verify the existing reels values to confirm no manufacturing flaws on part values, ie someone the reel changed resistance values on yesterday. EDIT ALL parts on the board seem to be correct with schematic.
3. Find an old AD628 in the graveyard box and stick it in and verify there is no mfg output change on newer devices. Makes no sense there would be mfg changes in the same tube of parts though. EDIT NO OLD PARTS FOUND to test.
4. Verify the input sense R's are correct at 1.2 ohm each. EDIT Done.
In the output of the AD628 amp, there is a divider section(R18, R21). Something would have to be really off on the math to show this much difference.
Any suggestions on how to diagnose this weirdness?
Thanks
pdf
16K
Comments
I love it when I find that I am not the only one to do such stupid things!
It's reminder of the old adage: Check your voltages first.
Often forgotten when we suspect so many other things.
-Phi;
Kudos to T Chap for your honesty and such a detailed spec of the problem. As you wrote, sometimes just taking a break from the bench and writing out the problem in a clear fashion is helpful to sorting out a problem. Good on ya. You're a pro and sometimes the simplest mistakes are the toughest ones to track down.